LLM EXPLORER 61,491 MODELS INDEXED

Brucethemoose Yi 34B 200K RPMerge 4.65bpw H6 EXL2 by MarinaraSpaghetti

By MarinaraSpaghetti · 8 downloads
HF Repo ↗ Copy URL

Brucethemoose Yi 34B 200K RPMerge 4.65bpw H6 EXL2 is an open-source language model by MarinaraSpaghetti. Features: 34b LLM, VRAM: 10.5GB, Context: 195K, License: other, Quantized, LLM Explorer Score: 0.11.

  Arxiv:2306.01708   Arxiv:2311.03099   En   Endpoints compatible   Exl2   Exllama   Exllamav2   Llama   Merge   Mergekit   Quantized   Region:us   Safetensors   Sharded   Tensorflow   Yi
Model Card on HF 🤗: https://huggingface.co/MarinaraSpaghetti/brucethemoose_Yi-34B-200K-RPMerge-4.65bpw-h6-exl2 

Brucethemoose Yi 34B 200K RPMerge 4.65bpw H6 EXL2 Parameters and Internals

Model Type 
text generation, instruct-enhanced storytelling
Use Cases 
Areas:
Storytelling, Instruct-enhanced tasks
Applications:
Long-context writing, Roleplaying scenarios
Primary Use Cases:
40K+ context instruct-enhanced storytelling
Limitations:
Some quirks due to token/instruct formats
Considerations:
Careful parameter tuning required for optimal performance.
Additional Notes 
This model is a merge of several Yi 34B models to achieve specific performance goals, especially for storytelling.
Supported Languages 
en (High)
Training Details 
Data Sources:
Roleplaying data, Vicuna-format data, Alpaca format data
Data Volume:
200K tokens
Methodology:
DARE TIES merge method
Context Length:
40000
Hardware Used:
24GB GPUs
Model Architecture:
Yi 34B with Vicuna-format emphasis
Input Output 
Input Format:
Orca-Vicuna
Accepted Modalities:
text
Output Format:
text
Performance Tips:
Use lower temperature settings and additional sampling techniques for best performance.
Release Notes 
Notes:
Initial release focusing on instruct-enhanced storytelling at 40K+ context.
LLM NameBrucethemoose Yi 34B 200K RPMerge 4.65bpw H6 EXL2
Repository 🤗https://huggingface.co/MarinaraSpaghetti/brucethemoose_Yi-34B-200K-RPMerge-4.65bpw-h6-exl2 
Model Size34b
Required VRAM10.5 GB
Updated2026-05-25
MaintainerMarinaraSpaghetti
Model Typellama
Model Files  8.6 GB: 1-of-2   1.9 GB: 2-of-2
Supported Languagesen
Quantization Typeexl2
Model ArchitectureLlamaForCausalLM
Licenseother
Context Length200000
Model Max Length200000
Transformers Version4.37.2
Tokenizer ClassLlamaTokenizer
Padding Token<unk>
Vocabulary Size64002
Torch Data Typebfloat16

Best Alternatives to Brucethemoose Yi 34B 200K RPMerge 4.65bpw H6 EXL2

Best Alternatives
Context / RAM
Downloads
Likes
...34B 200K Aezakmi Raw 1902 EXL2195K / 20.7 GB31
Yi 34B 200K MAR2024 EXL2 4bpw195K / 18 GB41
...mi Raw Toxic 2702 4.65bpw EXL2195K / 20.8 GB72
Yi 34B 200K RPMerge195K / 68.9 GB7260
Opus V1 34B 5.0bpw H6 EXL2195K / 22.3 GB52
Opus V1 34B 8.0bpw H8 EXL2195K / 34.8 GB71
Opus V1 34B 4.0bpw H6 EXL2195K / 18 GB51
34B Beta 5.0bpw H6 EXL2195K / 22.3 GB41
34B Beta 4.0bpw H6 EXL2195K / 18.1 GB21
Smaug 34B V0.1 2.65bpw H6 EXL2195K / 12.4 GB41
Note: green Score (e.g. "73.2") means that the model is better than MarinaraSpaghetti/brucethemoose_Yi-34B-200K-RPMerge-4.65bpw-h6-exl2.